This article examines the implication of the simple economic principle that price is determined by supply and demand in determining electric energy prices in deregulated US power markets. It questions whether electric energy prices can compensate non-utility generators and independent power producers for the cost building new generating capacity.

In examining this question, it is necessary to consider the very unique physical characteristics of electricity, since it cannot be easily stored, must be consumed at the same time that it is produced; and is delivered to end users by an interconnected electric grid where electricity follows the path of least resistance. These attributes require that installed electric capacity always be greater than system demand (or load); in other words, supply always exceeds demand. Concerns with system reliability dictate this need for excess capacity.¹

To address reliability issues, regulatory authorities commonly require a predetermined level of excess system capacity, or reserve margin, typically 15% of peak demand, defined as the greatest amount of electricity necessary to serve all customers in a given area at a given time. At the same time, actual electric demand varies significantly by season and by time of day. Average demand is significantly less than peak demand, and average daily off-peak demand is significantly less than on-peak demand. The consequence is that there is an enormous amount of unused generating capacity in all markets which must be paid for by ratepayers. As evidenced by numerous administrative proceedings at State public utility commissions and at the Federal Energy Regulatory Commission (FERC), including most recently the complaint by RPM Buyers against PJM Interconnection (PJM), a regional transmission organization (RTO) disputing PJM rules to determine certain capacity prices, ratepayers often contest the requirement to pay for this excess capacity.²

Complicating the price environment is the fact that retail electricity rates remain, for the most part, regulated under obsolete State-enforced obligation to serve rules imposed on load serving entities (LSEs) where "average retail rates disconnect the prices individual consumers pay from the marginal cost of providing them power in any hour"³ i.e. from the cost of wholesale generation.

The excess supply problem

Recall the basic economic premise that price will equal the marginal cost of production at market equilibrium. In electric markets, however, equilibrium is commonly thought of as the state where installed capacity equals peak demand plus a prescribed reserve margin. As discussed below, this has resulted in an approach to pricing that is based on supply-side engineering logic which is fundamentally different than economic logic that emphasizes demand-based pricing.

To understand the problem, consider the following in the context of a closed market, i.e. one that is not interconnected with other markets and therefore there are no imports or exports of power, with a 15% reserve margin. If each megawatt-hour (MWh) of electricity energy is sold at a price that is equal to the full cost per MWh for each generator (inclusive of depreciation and interest expense), system revenues will never be sufficient to permit full cost recovery of capital investment, since capacity utilisation will be low relative to total installed capacity. Put another way, the cumulative payments will not reimburse generators for supplying the capacity necessary to maintain the reserve margin. Therefore, full cost recovery can only be achieved through other means, principally higher prices in periods of higher demand or capacity payments.

Wholesale electricity prices

Deregulation of wholesale electric power markets in the US began in 1992 with the National Energy Policy Act, which allowed power producers to compete for the sale of electricity to utilities, and was further expanded when FERC issued Order 888 requiring utilities to open their transmission lines to competitors. Because implementation of deregulation has largely been enforced at the State level, market design and methods to establish wholesale prices vary by region. To understand common underlying pricing principles, it is instructive to look at the methodology of pricing models used by market participants.

The basic methodology for forecasting electric energy prices estimates (i) future load to determine the amount of required electric production, (ii) new capacity additions and retirements, and (iii) future fuel prices to establish the competitive position of different resources. Electric facilities are then dispatched in merit order, i.e., according to their marginal variable cost of production, taking into account fuel choice, variable operation and maintenance (O&M) expenses, and individual plant characteristics such as heat rate and air emissions. An electric price is then estimated to be the highest price necessary to clear the market, i.e., to provide sufficient electricity to satisfy forecast demand for a given time in a given region. Models vary in sophistication with respect to the methodology of incorporating other factors, such as time-of-use (typically projecting hourly demand and the hourly marginal cost of production, sometimes aggregated into on-peak and off-peak periods), transmission constraints, e.g., congestion, disruptions and line losses, and unplanned plant outages.

Such production cost models are supplemented by capacity valuation models to determine the full price (or value#4) of electricity, which is sometimes presented as distinct prices for both electric energy and capacity, and sometimes presented as a single energy price. The capacity price, whether embedded in an energy price or presented as a distinct variable, is in turn a function of the revenues required to support the cost of new entry (commonly referred to as CONE). The underlying, supply-oriented logic of this methodology presumes that new capacity will not be built unless future revenues received for electric power is sufficient to fully compensate the owners of a new facility for their total costs of construction and operation, including a return on invested capital, i.e., that price will be equivalent to the marginal cost of energy plus an amount equal to the reliability value-added by a new generator. It is this logic, which is derived from a regulated utility model and which is incorporated into market forecasting models, that is inconsistent with economic principles.

For example, one would presume that an airline, in placing a new aircraft order, does not assume ticket prices will increase to a level sufficient to pay for the new planes. Instead the airline evaluates expected demand at different price and different service levels and seeks to make investments and manage costs to provide various levels of acceptable service to attract customers and earn profits. Airlines, as opposed to LSEs, do not have an obligation to serve.

Economic dispatch and energy prices

Generally, electric energy prices are determined in a manner consistent with the assumptions of production cost models, i.e., a system administrator receives bids and dispatches generators from lowest to highest bids until output satisfies system demand. Unlike in the production cost models, however, the process of merit order dispatch does not necessarily take place in efficient markets and is interrupted by a number of real world factors, or non-market mechanisms, that conflict with the efficient market assumptions in forecasting models. Such non-market mechanisms, many of which are carried over from the old regulated regime, include: wholesale market price caps, capacity obligations placed on LSEs, frequency regulation, operating reserve and other ancillary service requirements, and protocols for non-price rationing of demand to respond to shortages.#6

Among the more significant interruptions are unplanned plant outages and transmission disruptions. One study, for example, noted that in New England (around 2004) "one quarter of fossil fuel production capacity is routinely utilized 'out of merit order' ... primarily as a result of unplanned outages."7# The extent of out-of-merit dispatch is evident in the number of facilities operated under reliability-must-run (RMR) contracts. For example, in California, over 80 plants amounting to in excess of 9,500MW were approved to operate under 2006 RMR contracts.#8

Perhaps more problematic is the antiquated US transmission grid. While transmission improvements have been significant in recent years, after a prolonged period of underinvestment, according to Governor and former Secretary of Energy Bill Richardson, "We still have a third-world grid."#9 At the same time Federal and State policies provide significant incentives to build wind and other renewable projects even though wind is not a well suited resource with regard to system reliability. While grid constraints limit the development of wind energy, the variability of wind resources also materially complicates dispatch decisions and grid management. The US grid is also systematically disturbed by common weather events.

Fixed-cost compensation and capacity markets

Since deregulation of power markets began in the 1990's, there have been ongoing efforts in many regional markets to establish price signals and to compensate wholesale generators for new capacity additions without sending prices to intolerable levels.#10

The general assumption in market forecast models is that generators will be paid for maintaining adequate capacity to meet peak load and it is therefore rational to include the capacity values in market price forecasts. This may be true in relatively competitive and liquid power markets#,11 but it has been demonstrated that this assumption fails in situations of overcapacity (although it is less well-established how overcapacity should be defined).

Essentially, as noted above, capacity markets are necessitated by the need for system reliability. As observed by S&P: "Absent fully unconstrained prices at the retail and wholesale levels, an efficient energy market alone cannot sufficiently incent long-term investment. As electric energy is an essential commodity and because the economic consequences of running out are simply unacceptable, market rules have been put in place to create a market for generation capacity. This separate capacity market is to ensure that generators have the incentive to keep adequate generation available.#"12

Formal capacity markets now exist in three regions: PJM, NY ISO (Independent System Operator), and NE-ISO. While each regime is different, each "serves to recruit electric-generating capacity three years ahead of time, in order to keep the lights on."13# In the case of PJM and NE-ISO, the auctions establish three-year forward capacity prices, which allow new entrants sufficient time to build new facilities. Similarly, NY ISO determines capacity prices according to results of several auctions (semi-annual, monthly and spot) using administratively determined demand curves (one for each sub-region), which are reviewed every three years to make necessary adjustments.

Table 1 provides a history of recent capacity prices determined through auction for PJM and NY ISO, as well as transition rates (approved in a settlement proceeding) for NE-ISO. As can be seen, recent capacity prices are not really consistent with current CONE estimates of approximately $7.50 to $9.00 per kW-month. At the same time, objections to these rates are evident in the decision of Duquesne Light to withdraw from PJM, the RPM buyers' complaint, and the consideration by Maine to withdraw from NE-ISO are disconcerting reactions to the costs approved to date through the competitive capacity proceedings.

Conclusion

Although there are persuasive arguments and valid theory (not to mention thousands of pages of supporting testimony in various proceedings) underlying the utility and economic effectiveness of capacity markets, the current state of progress is not all that encouraging. It can be reasonably argued that the need to finance, develop and build new electric generating capacity requires perseverance in implementing a market framework and regulatory policies and procedures to reinforce deregulation and support the evolution of capacity markets. However it should also be acknowledged that "the combination of the unusual physical attributes of electricity and electric power networks and associated reliability considerations, limitations on metering of real time consumer demand and responsiveness of real time prices, restrictions on the ability to ration individual consumers, discretionary behaviour by system operators, makes achieving an efficient allocation of resources with competitive wholesale and retail market mechanisms a very challenging task."#14

The challenge is further complicated by policy considerations with respect to fuel diversity and carbon emissions, renewable portfolio standards, and expansion and improvement of the national electric transmission grid.

In consideration of these varied challenges, particularly with respect to the disconnect between retail and wholesale markets, it does not seem reasonable to believe that current capacity markets will adequately compensate owners and investors in new electric generating capacity unless there is a better coordination of policy objectives and a more coherent national energy policy.

Richard Garcia is a senior managing director and Head of Project Finance for the Americas, Natixis New York Branch. The opinions expressed in this article are those of its author and may differ from those of Natixis.

Footnotes
1 See Shmuel S. Oren, "Capacity Payments and Supply Adequacy in Competitive Electricity Markets," paper presented at the VII Simpósio de Especialistas em Planejamento da Operação e Expansão Eletrica, May 2000. Oren explains that reliability encompasses two attributes of electric systems: (i) security, which is the ability of the system to withstand disturbances; and (ii) adequacy, which represents a system's ability to meet power requirements of all consumers at all times.
2 Maryland PUC, et al. vs. PJM Interconnection, LLC; FERC Docket EL08-67. Note that RPM refers to the Reliability Pricing Model. See also Bruce W. Radford, "Buyer's Remorse: The PJM complaint and the rising cost of electric reliability," Fortnightly, September 2008.
3 Lynne Kiesling, "Market-based Electricity Pricing," Reason Public Policy Website.
4 Capacity value is essentially the value of system reliability, commonly measured as the value of lost load (VOLL).
5 Perhaps an analogous activity is the hospital industry, where pricing practices also present public policy concerns and debate.
6 Paul Joskow and Jean Tirole, "Reliability and Competitive Electricity Markets," April 21, 2004 (page 2).
7 William Klun, "Investment Risk and the Economics of Capacity," Power Finance and Risk, July 16, 2004. As observed by Josko and Tirole, op. cit., when a less efficient unit is dispatched out-of-merit, it is treated as a resource with a $0 offer, thereby leading to a decrease in the market price (page 28).
8 Per CAISO 2006 RMR Contract Status list updated on March 21, 2006.
9 Matthew L. Wald, "Wind Energy Bumps into Power Grid's Limits," New York Times, August 27, 2008.
10 These efforts have been severely tested by incidents such as the California energy crisis in 2001 – '02, and there is no shortage of critics who have concluded that deregulation has been a failure; see, e.g. Marilyn Showalter, "Deregulation Means Higher Prices," EnergyBiz online, November/December 2007.
11 Within the US there are five relatively liquid markets, i.e., PJM, NY ISO, ERCOT, CAISO and NE-ISO. Other organized markets, i.e., where a regional transmission organization (RTO) or independent system operator (ISO) serves as a third-party independent operator, include the Midwest ISO and Southwest Power Pool (SPP). The Electric Power Supply Association (EPSA) estimates that two-thirds of the United States' economic activity occurs within the boundaries of these markets. See: http://www.epsa.org/industry/primer/?fa=rto.
12 S&P Ratings Direct Report "Proposed New Long-Term Capacity Markets in the Eastern US Are Favorable for Electric Generators," September 15, 2006.
13 Bruce W. Radford, "Buyer's Remorse" op. cit.
14 Joskow and Tirole, op. cit. (page 48).